Decomposition and delexicalisation in learners’ collocational (mis)behaviour.

This paper presents an ongoing study into the cognitive and linguistic aspects of lexical collocation, and how these relate to collocation errors in a corpus of learner English compiled by the author. Although there is a growing body of research based on such data, very little attention is generally paid to the cognitive aspects which underlie lexical and collocational acquisition. For native speakers, meaning by collocation involves some degree of delexicalisation. Delexicalisation colludes in masking the arbitrariness of many collocations by detracting the speaker’s attention from the fully salient meanings of the collocates: highly original expresses the extent of originality not its height. Learners, however, are more inclined to favour literal, compositional meaning, both in comprehension and production. When phraseological chunks are broken down into their component parts, there is a serious risk of their meaning being distorted, typically through over-literal interpretation. L2 collocations that do not correspond to L1 patterns are frequently avoided in learner writing, with calqued L1 collocations taking their place. These show up as collocation errors because, being unconventional, they are read compositionally, and hence as semantically anomalous. This paper discusses the pedagogical importance of comparing collocational patternings in L1 and L2, the need to account more satisfactorily for phraseological meanings in the vocabulary acquisition process. It argues the case in favour of developing more effective approaches to the teaching and learning of conventionalised, delexical language, in which collocational motiviation is revisited from the non-native speaker’s point of view.